Recording and Mixing Levels in Tracktion
-
- KVRist
- 287 posts since 15 May, 2005
I'm an inexperienced user of Tracktion ( still on version 2) who has got confused by some stuff I've been reading about recording and mixing levels.
I've always recorded audio and midi into T using the track level meters to make sure that no peaks occurred above 0db. I know what digital clippping can sound like on vocals - not nice.
The general guidance seemed to be to aim for as loud a signal when recording, without breaching 0db, so as to maximize signal to noise ratio.
I've read things in a few places on the web recently that says when recording in the digital domain, its best to achieve peaks no higher than -12db, and the average level will probably be around -18db,so as to give plenty of headroom for possible future adjustments e.g. eq boosts, and for mastering. It also suggests that the master output should be set to 0db when recording and mixing, and track levels adjusted accordingly.
However, some of the instrument plug-ins I use don't seem to have enough volume range in them to be usable at such low recording levels.
I've never got to the point of transferring a 'finished' song out of Tracktion onto CD or mp3, and have never tried anything on the mastering side.
When mixing (if I can call it that!) I do basic level setting of tracks, panning, effects and eq, with the master output left at its default -3db.
When I have needed to boost eq on a track (even a 12db boost) it has not led to clipping even if the track and master level meters go into the red. However, I tend to cut eq, rather than boost if possible, as this seems 'best practice'.
I do have the luxury of being able to rerecord things to make them fit better in the mix, rather than having to 'fix it in the mix'.
However, I've seen experienced Tracktion users on this forum talk about having the master output level at anything like -16db when they've finished mixing, as that is what results from the way they have adjusted individual tracks.
Is this better because it gives more headroom for mastering, than having the master level set at 0db when mixing?
I'd really appreciate comments from you experienced Tracktion users on the 'best' approach as I'm finding this confusing!
I don't want to have recorded a lot of songs, and when it comes to proper mixing and maybe mastering e.g. to go onto a CD, I find I've recorded things in a way that leads to problems at later stages in the process.
Hope this makes sense!
Thanks
I've always recorded audio and midi into T using the track level meters to make sure that no peaks occurred above 0db. I know what digital clippping can sound like on vocals - not nice.
The general guidance seemed to be to aim for as loud a signal when recording, without breaching 0db, so as to maximize signal to noise ratio.
I've read things in a few places on the web recently that says when recording in the digital domain, its best to achieve peaks no higher than -12db, and the average level will probably be around -18db,so as to give plenty of headroom for possible future adjustments e.g. eq boosts, and for mastering. It also suggests that the master output should be set to 0db when recording and mixing, and track levels adjusted accordingly.
However, some of the instrument plug-ins I use don't seem to have enough volume range in them to be usable at such low recording levels.
I've never got to the point of transferring a 'finished' song out of Tracktion onto CD or mp3, and have never tried anything on the mastering side.
When mixing (if I can call it that!) I do basic level setting of tracks, panning, effects and eq, with the master output left at its default -3db.
When I have needed to boost eq on a track (even a 12db boost) it has not led to clipping even if the track and master level meters go into the red. However, I tend to cut eq, rather than boost if possible, as this seems 'best practice'.
I do have the luxury of being able to rerecord things to make them fit better in the mix, rather than having to 'fix it in the mix'.
However, I've seen experienced Tracktion users on this forum talk about having the master output level at anything like -16db when they've finished mixing, as that is what results from the way they have adjusted individual tracks.
Is this better because it gives more headroom for mastering, than having the master level set at 0db when mixing?
I'd really appreciate comments from you experienced Tracktion users on the 'best' approach as I'm finding this confusing!
I don't want to have recorded a lot of songs, and when it comes to proper mixing and maybe mastering e.g. to go onto a CD, I find I've recorded things in a way that leads to problems at later stages in the process.
Hope this makes sense!
Thanks
-
- KVRist
- 260 posts since 15 Nov, 2006
I usually end up with -5 to -10 on tracks and often lower. If levels are much higher than this then when tracks are grouped together on a buss then the sum level will end up to high. Also, as you said, you need some room for FX etc. You don't wan't any overs on the tracks or the master buss at all but if you end up with a nice mix that ends up being high on the master buss (but not over say -1 peak) then you can just lower the level to -10 or-12 when you render it to stereo. This will leave plenty of room for the mastering eng. who works only on the stereo file after it's mixed. If you try to master it yourself then treat the mixing and mastering separately rather than sticking mastering FX on the stereo buss.
Steve
-
- KVRist
- 139 posts since 17 May, 2006
This is great advice that, as a home studio recordist, took me about 2 years to figure out on my own...ivegotmono wrote: If you try to master it yourself then treat the mixing and mastering separately rather than sticking mastering FX on the stereo buss.
-------
The pump don't work cuz the vandals took the handle!
The pump don't work cuz the vandals took the handle!
-
- KVRist
- Topic Starter
- 287 posts since 15 May, 2005
Thanks for your replies. This confirms what I suspected, that headroom needs to be left for later stages in the process.
I've never done a grouping of instruments or vocals together, but I can see that doing so would require the individual track levels of, for example, 3 ltracks of backing vocals, when grouped together, to be lowered.
But, presumably, there would be no problem in doing that, even if I had recorded each of the 3 individual backing vocal tracks to peak at between 0db and -3db. I would simply lower the volume levels for mixing.
It seems that individual track levels are going to constantly creep down as mixing takes place.
What I don't understand is why recording levels should be at -12db peak and approximately -18db average.
Is that the way you guys record?
Thanks
I've never done a grouping of instruments or vocals together, but I can see that doing so would require the individual track levels of, for example, 3 ltracks of backing vocals, when grouped together, to be lowered.
But, presumably, there would be no problem in doing that, even if I had recorded each of the 3 individual backing vocal tracks to peak at between 0db and -3db. I would simply lower the volume levels for mixing.
It seems that individual track levels are going to constantly creep down as mixing takes place.
What I don't understand is why recording levels should be at -12db peak and approximately -18db average.
Is that the way you guys record?
Thanks
-
- KVRist
- 260 posts since 15 Nov, 2006
I usually start out with recorded levels of -5to-10 sometimes a little higher and very often lower. I haven't had any problems unless I get down to -20 or so and then the noise does come into play.
Steve
-
- KVRAF
- 4644 posts since 28 Nov, 2002 from Chicago
hmmm...
Some of the above is not quite true.
Back in the day, with digital it was necessary to record pretty hot. As you start dropping below 16 bits, the quslity drop becomes noticeable quickly.
The flip side is that above 16bits, each extra bit makes exponentially less difference to perceived audio quality. You might be able to tell the difference between 16 and 24 bit, but between 24 and 32, not so much. Similarly, the difference between 16 and 20 is much greater than, say 20 and 24. What that means is that with 24bit converters you can safely leave generous amounts of headroom.
Even -12dB is still way above 16 bits, and for the most part, your final mix will be at 16 bits.
Why leave headroom? Simple. You can't rely on live performers not to generate occasional peak levels that are way higher than their average. People used to throw compressors in front of digital converters to control these peaks. These days we don't need to, so we can tailor compression at leisure once the recoring is made.
That's it, that's all there is. Low levels is simply about not clipping the inputs, while making use of the *huge* dynamic range of 24bit converters.
Regards leaving space for post processing, this is only relevant for some outboard gear that does its calculations in integer space. Almost all DAWS use floating point, so there is no headroom to preserve. Once the audio has passed through the ADC, clipping is a null issue. With the exception of a few badly programmed plugins, the only clipping that can occur once audio is recorded is at the output stage.
You can ignore your track meters, they can run way over redline, it doesn't matter at all. You can't ignore the master meter though. If you peak above 0dBs there, you'll hear it.
So simple rule: don't clip on the way in, don't clip on the way out, don't worry about the bit in the middle.
Some of the above is not quite true.
Back in the day, with digital it was necessary to record pretty hot. As you start dropping below 16 bits, the quslity drop becomes noticeable quickly.
The flip side is that above 16bits, each extra bit makes exponentially less difference to perceived audio quality. You might be able to tell the difference between 16 and 24 bit, but between 24 and 32, not so much. Similarly, the difference between 16 and 20 is much greater than, say 20 and 24. What that means is that with 24bit converters you can safely leave generous amounts of headroom.
Even -12dB is still way above 16 bits, and for the most part, your final mix will be at 16 bits.
Why leave headroom? Simple. You can't rely on live performers not to generate occasional peak levels that are way higher than their average. People used to throw compressors in front of digital converters to control these peaks. These days we don't need to, so we can tailor compression at leisure once the recoring is made.
That's it, that's all there is. Low levels is simply about not clipping the inputs, while making use of the *huge* dynamic range of 24bit converters.
Regards leaving space for post processing, this is only relevant for some outboard gear that does its calculations in integer space. Almost all DAWS use floating point, so there is no headroom to preserve. Once the audio has passed through the ADC, clipping is a null issue. With the exception of a few badly programmed plugins, the only clipping that can occur once audio is recorded is at the output stage.
You can ignore your track meters, they can run way over redline, it doesn't matter at all. You can't ignore the master meter though. If you peak above 0dBs there, you'll hear it.
So simple rule: don't clip on the way in, don't clip on the way out, don't worry about the bit in the middle.
Someone shot the food. Remember: don't shoot food!
-
- KVRist
- 445 posts since 31 Dec, 2005 from Albuquerque
This is what I've always thought as well. It's good to have some confirmation.valley wrote:hmmm...
Some of the above is not quite true.
Back in the day, with digital it was necessary to record pretty hot. As you start dropping below 16 bits, the quslity drop becomes noticeable quickly.
The flip side is that above 16bits, each extra bit makes exponentially less difference to perceived audio quality. You might be able to tell the difference between 16 and 24 bit, but between 24 and 32, not so much. Similarly, the difference between 16 and 20 is much greater than, say 20 and 24. What that means is that with 24bit converters you can safely leave generous amounts of headroom.
Even -12dB is still way above 16 bits, and for the most part, your final mix will be at 16 bits.
Why leave headroom? Simple. You can't rely on live performers not to generate occasional peak levels that are way higher than their average. People used to throw compressors in front of digital converters to control these peaks. These days we don't need to, so we can tailor compression at leisure once the recoring is made.
That's it, that's all there is. Low levels is simply about not clipping the inputs, while making use of the *huge* dynamic range of 24bit converters.
Regards leaving space for post processing, this is only relevant for some outboard gear that does its calculations in integer space. Almost all DAWS use floating point, so there is no headroom to preserve. Once the audio has passed through the ADC, clipping is a null issue. With the exception of a few badly programmed plugins, the only clipping that can occur once audio is recorded is at the output stage.
You can ignore your track meters, they can run way over redline, it doesn't matter at all. You can't ignore the master meter though. If you peak above 0dBs there, you'll hear it.
So simple rule: don't clip on the way in, don't clip on the way out, don't worry about the bit in the middle.
-
- KVRist
- 295 posts since 27 Nov, 2006 from Limoges/France
Very interesting post.
So if I get what Valley says : if I have one track with a "natural" (ie recorded) volume between say -5db and 0db, and add some filters on it and it ends way above 0db, I will not loose any sound quality lowering it with a volume filter.
So if I get what Valley says : if I have one track with a "natural" (ie recorded) volume between say -5db and 0db, and add some filters on it and it ends way above 0db, I will not loose any sound quality lowering it with a volume filter.
-
- KVRist
- 131 posts since 12 Jan, 2007 from Victoria, B.C., Canada
I don't think you have it...no. 'Natural' recording in current digital is -18dbRMS (average) with peaks -9db or so. Some folks even record at -24dbRMS or less -- there is no fear of quality loss. Keep your headroom. You will never regret it.Very interesting post.
So if I get what Valley says : if I have one track with a "natural" (ie recorded) volume between say -5db and 0db, and add some filters on it and it ends way above 0db, I will not loose any sound quality lowering it with a volume filter.
Check out this very helpful link on Tweak's website re levels. It's long, but worthwhile reading.
http://studio-central.com/phpbb/viewtopic.php?t=34078
A.
- KVRAF
- 1855 posts since 21 Sep, 2004 from Musician, Recording Engineer, Producer
You got it. Don't worry about leaving headroom at the recording stage, just worry about clipping on the input. For the second reason, don't worry about recording too hot. It's much better to record it a bit lower in volume and avoid clipping. A good rule of thumb that I heard is that on melodic sources (non percussion) when zoomed out the waveform should take up at least %50 of the trackspace on average. Too low and you loose resolution, too high and you risk clipping. Somewhere in the middle is the common ground.Njxt2 wrote:Very interesting post.
So if I get what Valley says : if I have one track with a "natural" (ie recorded) volume between say -5db and 0db, and add some filters on it and it ends way above 0db, I will not loose any sound quality lowering it with a volume filter.
Oh... and another thing... If you record a great take and the meters show clipping, but you don't hear any clipping artifacts, keep it. Listen really hard. If it just clipped for a few samples it is sometimes inaudible, and will work fine in a mix. Listen carefully though.
- KVRAF
- 3462 posts since 17 Sep, 2006 from Fredericksburg, VA USA
I read alot of this excellent article and now try to keep my levels around -15DBFS. I find that the end product now has a much greater dynamic range. Also, it is my understanding that -15DBFS is the quuivalent of 0dbvu - please correct me if I'm wrong.tonyh wrote:I don't think you have it...no. 'Natural' recording in current digital is -18dbRMS (average) with peaks -9db or so. Some folks even record at -24dbRMS or less -- there is no fear of quality loss. Keep your headroom. You will never regret it.Very interesting post.
So if I get what Valley says : if I have one track with a "natural" (ie recorded) volume between say -5db and 0db, and add some filters on it and it ends way above 0db, I will not loose any sound quality lowering it with a volume filter.
Check out this very helpful link on Tweak's website re levels. It's long, but worthwhile reading.
http://studio-central.com/phpbb/viewtopic.php?t=34078
A.
D Scarlatti, Dell XPS8700 i7/8gb mem/1tb hd/Steiny UR22/Presonus ER5s/Nektar LX61 kbd ctrlr/Win 10 Pro/S1 4.6/ my music here: https://www.magix.info/us/profile/my-profile/media/
-
- KVRist
- 131 posts since 12 Jan, 2007 from Victoria, B.C., Canada
I understand it to be -18 dbfs, but again you have so much freakin' room with current digital DAW's that there really is no point in quibbling about a few dbf's. A quote from John Scrip of Massive Mastering:
"A signal that *PEAKS* at -47dBFS still has higher resolution than a compact disc. When you think about it that way, understanding that there is *so much* headroom in 24-bit - way more than anyone's gear can handle, making a fuss over getting a lousy couple dB louder really seems silly, no?"
"A signal that *PEAKS* at -47dBFS still has higher resolution than a compact disc. When you think about it that way, understanding that there is *so much* headroom in 24-bit - way more than anyone's gear can handle, making a fuss over getting a lousy couple dB louder really seems silly, no?"
-
- KVRAF
- 4644 posts since 28 Nov, 2002 from Chicago
I don't think you're quite understanding his question.tonyh wrote:I don't think you have it...no. 'Natural' recording in current digital is -18dbRMS (average) with peaks -9db or so. Some folks even record at -24dbRMS or less -- there is no fear of quality loss. Keep your headroom. You will never regret it.Very interesting post.
So if I get what Valley says : if I have one track with a "natural" (ie recorded) volume between say -5db and 0db, and add some filters on it and it ends way above 0db, I will not loose any sound quality lowering it with a volume filter.
Check out this very helpful link on Tweak's website re levels. It's long, but worthwhile reading.
http://studio-central.com/phpbb/viewtopic.php?t=34078
A.
Once the audio is safely *inside* Tracktion, only badly behaved plugins will care if it goes above 0dBs.
So yeah, you can run a track at over 0dBS, and pull the levels down elsewhere, if for some reason you want to.
Someone shot the food. Remember: don't shoot food!
-
- KVRist
- 295 posts since 27 Nov, 2006 from Limoges/France

